My research interests broadly lie in Quantum Information, Network Coding, Machine Learning, Artificial Intelligence in Healthcare and Genomics, and Data Compression, with a strong emphasis on rigorous theoretical foundations and impactful interdisciplinary applications.

My core research is grounded in information theory and network coding, inspired by the seminal work of Claude E. Shannon, the Father of Information Theory, who established fundamental limits on data compression, reliable storage, and communication. Central to this framework are Shannon’s four information measures—entropy, conditional entropy, mutual information, and conditional mutual information—and the associated information inequalities that characterize feasible regions in entropy space. A fundamental open problem in information theory is determining whether a given point in entropy space is entropic, with the collection of all such points forming the entropic region. Characterizing this region is crucial for solving deep problems in network coding, such as determining feasible transmission rates and code constructions.

My doctoral research, supervised by Dr. Satyajit Thakor, focused on the characterization of entropy regions and their applications to network coding. In particular, I studied inner and outer bounds of the entropic and almost-entropic regions and investigated optimization problems over these regions, including questions related to the Ingleton inequality and the Four-Atom Conjecture. These results provide theoretical foundations for determining the capacity (rate) regions of multi-source multi-sink network coding problems, where classical store-and-forward strategies are suboptimal and network coding offers significant gains in throughput, security, and efficiency. Parts of this work have been published in IEEE Transactions on Communications, ISITA, ISCON, and arXiv.

Building on this strong theoretical base, my recent research extends information-theoretic principles to modern data science and AI systems, particularly focusing on model robustness, interpretability, and generalization. In this direction, I served as the Principal Investigator of a sanctioned international research project funded by OpenAI Inc., USA, titled “Leveraging Information Theory Metrics to Enhance Predictive Accuracy, Interpretability, and Robustness of AI Models in Diverse Data Science Applications”. This project (US $5,000; Aug 2024–Apr 2025) investigated the use of entropy-based and mutual-information-driven metrics to analyze and improve learning dynamics in AI models, bridging classical information theory with contemporary machine learning and explainable AI frameworks.

Prior to my Ph.D., I worked as a Junior Research Fellow under Dr. Syed Abbas on an NBHM-funded project titled “Periodicity and Almost Periodicity in Ecological Modelling” at IIT Mandi. This work aligns with my broader interest in mathematical modeling of complex systems. Using time-scale calculus, which unifies continuous and discrete analysis, we studied predator–prey dynamics via a modified Leslie–Gower model with Holling-type II functional response and Michaelis–Menten harvesting, deriving verifiable conditions for periodic solutions using coincidence degree theory.

My interdisciplinary research journey began during my M.Tech project, where I applied information-theoretic and computational modeling techniques to genomics. I analyzed LINEs and SINEs retrotransposons in Entamoeba histolytica using the quasispecies model, estimating mutation rates, studying sequence polymorphism, and constructing phylogenetic trees using datasets from Prof. Sudha Bhattacharya’s laboratory at JNU. This work highlighted the high mutation rates and evolutionary dynamics of SINE elements and laid the foundation for my ongoing interest in AI-driven genomic analysis, data compression, and healthcare applications.

Overall, my research integrates deep theoretical advances in information theory and network coding with emerging challenges in AI, data science, and genomics, aiming to develop mathematically principled, robust, and interpretable computational models for next-generation communication and intelligent systems.